171 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
ZacharyZhang-NY	f7e5563a71	T48: round 2 fix: at a new DPI her corner follows once the move that brought it is done, and only where her box would lie mostly on a display of that DPI; else Windows' rect stands, since between two displays side by side a corner kept always set her DPI back and forth; the tests move beside the module	2026-10-18 13:16:13 -04:00
ZacharyZhang-NY	8ddf470f73	T48: round 1 fix: her height is her own DPI's and her bottom the pane's top whatever the two windows' DPI, where the main window's scale had put her 62 pixels into the terminal or above it; and when a move takes her to a display of another DPI, gpui takes the rect Windows suggests while she keeps her bottom left corner	2026-10-18 12:55:30 -04:00
ZacharyZhang-NY	c1cdddf07e	T48: Ely on Windows in a box her own size, as a user found her broken in 0.1.4: gpui opened her popup as large as the display when the middle of the box asked for lay off it, as an overlapped window Windows framed, so she sat at its bottom over the terminal and it took the clicks; her window becomes a bare popup, and the glue posts it her size, from her own DPI, with her place on the pane's top edge, since gpui hears a new size only outside its own updates	2026-10-18 12:43:25 -04:00
ZacharyZhang-NY	5fd4f8faa3	T47: reviewed, ticked	2026-10-18 05:58:45 -04:00
ZacharyZhang-NY	cdafdc2836	T47: release v0.1.4 at 2d9180e, its three builds from their scripts, the zip on the owner's Windows VM, the tarball on the owner's PC under Omarchy, the disk image signed and notarized by the Release workflow, published once CI passed on all three	2026-10-18 05:51:40 -04:00
ZacharyZhang-NY	2d9180e777	T47: version 0.1.4	2026-10-18 05:34:04 -04:00
ZacharyZhang-NY	1daa2599f7	T46: reviewed, ticked	2026-10-18 05:33:15 -04:00
ZacharyZhang-NY	6df7534e7f	T46: round 1 fixes: Ely's window never takes the keys, a panel key only if needed on macOS and on Windows a window that answers a press's activation with none; she opens with the sessions' mood; she stands 140 points in or less on a narrow pane, so she holds its edge at any width, and on Wayland, whose popup keeps the place it opened at, she opens again when a new width moves her	2026-10-18 05:26:13 -04:00
ZacharyZhang-NY	9aeb3f3552	T46: Ely peeks over the main window's top edge from behind it, as the owner asked after the 0.1.3 film: her head and fingers from the icon's paths in a transparent window of her own, kept on the edge as each platform allows, macOS's child window, Windows' owned popup, X11's popup moved and stacked above the frame, Wayland's anchored popup; ten faces follow the most pressing of the sessions' marks, with hops, a shake and a rise; a press winks; Settings, Terminal, Ely turns her off	2026-10-18 05:06:59 -04:00
ZacharyZhang-NY	b35eb52d9a	T45: reviewed, ticked	2026-10-18 04:58:13 -04:00
ZacharyZhang-NY	c30bd51fba	T45: round 2 fix: while herdr takes the mouse every link press is herdr's again, its rules with it, as its client and docs say: it keeps its panes' OSC 8 targets from the terminal so that it alone answers Ctrl-press, and opens web links only, so the terminal opening a file:// label it sees opened the wrong target; Shift with the press, herdr's named way past it, lets the terminal open what it sees, a file:// link too	2026-10-18 04:54:55 -04:00
ZacharyZhang-NY	af66a70947	T45: round 1 fix: herdr's client opens only http and https links, so inside herdr a link the app finds of another kind, file:// among them, the app opens itself; web links and links the app cannot see still go to herdr	2026-10-18 04:46:06 -04:00
ZacharyZhang-NY	ae3de244c2	T45: Cmd-press, Ctrl-press off macOS, opens links inside herdr too: herdr wraps a long URL at its pane's edge and passes no OSC 8 target on, so the app opened a cut URL or nothing; while herdr's client leads the session and tracks the mouse the press goes to herdr, which opens its pane's link whole and underlines it under the key, Cmd heard as Ctrl on macOS; elsewhere the app opens links as before	2026-10-18 04:35:49 -04:00
ZacharyZhang-NY	6b2ec654dc	T44: reviewed, ticked	2026-10-18 04:31:43 -04:00
ZacharyZhang-NY	3f47134928	T44: round 2 fix: on Windows a repeat goes as a record where its press did, one that types after its modifiers came up too, so its release still goes as a record and herdr frees the key; the choice is one function, records, tested on every system	2026-10-18 04:28:55 -04:00
ZacharyZhang-NY	6f8bfaeff8	T44: round 1 fixes: under kitty's flags the legacy bytes a key falls back to keep gpui's character, so Option-Shift-[ with only events set sends ESC { again; a key that is down is known by its key code on macOS and its virtual key on Windows, so its release goes out, as a record where its press did, though Shift came up first; shot posts a press or a release alone with down: and up:	2026-10-18 04:21:52 -04:00
ZacharyZhang-NY	fc10c4d491	T44: herdr's shortcuts work, which a user found dead: under kitty's flags a key Shift makes a symbol of goes as its own key with Shift, the symbol its alternate, as Ghostty sends it, read from the layout gpui reads on macOS and Windows; on Windows keys that type no text go as win32-input-mode records, which ConPTY asks for and herdr reads, as Windows Terminal sends them; on Linux gpui keeps no key code, which the docs say; shot learns wait, n and ]	2026-10-18 04:00:24 -04:00
ZacharyZhang-NY	8d7e730a78	Plan: T44 herdr's shortcuts on every system, T45 links open on Cmd-press and Ctrl-press, T46 Ely peeks over the window with a switch to turn her off, as the owner asked; T47 release v0.1.4	2026-10-18 03:24:06 -04:00
ZacharyZhang-NY	3ee525a008	T43: v0.1.3 rebuilt with T42, as the owner asked in place of a v0.1.4: the three builds from eb96109 replace the first ones on the release, each downloaded back matches the new SHA-256 in the notes, and the tag and the release's target moved to eb96109; no review, as the owner asked; ticked	2026-10-17 23:17:08 -04:00
ZacharyZhang-NY	eb9610988e	T42: a session's status mark moves, as the owner asked after aicss.dev's orbs: 3 by 3 dots in the state's color; work ripples from the centre, progress fills the outer ring clockwise from the top, waiting stays lit and dips twice, error, done and idle draw an X, a check and a ring, then hold; the window draws at the marks' 80 ms steps only while one moves; no review, as the owner asked v0.1.3	2026-10-17 22:38:46 -04:00
ZacharyZhang-NY	1e08cae4c4	T41: reviewed, ticked	2026-10-17 16:46:41 -04:00
ZacharyZhang-NY	7569f0011f	T41: version 0.1.3	2026-10-17 16:23:13 -04:00
ZacharyZhang-NY	336c05beb7	T40: reviewed, ticked	2026-10-17 16:22:44 -04:00
ZacharyZhang-NY	c62c17fa7c	T39: reviewed, ticked	2026-10-17 16:18:47 -04:00
ZacharyZhang-NY	140c2a5741	T40: review round 2: the session's title comes from the tap alone, in the output's order, so the row, the window and a notification name it alike: OSC 0 and 2 read as vte splits them, 15 parts past the number, and up to 65536 bytes; alacritty's title events, which come later, pass by, and a full reset no longer sets the options again to send one	2026-10-17 16:17:55 -04:00
ZacharyZhang-NY	22752a28db	T40: review round 1: a notification names the session as the output stands at its place: the tap follows OSC 0 and 2, the title stack's CSI 22 t and 23 t and a full reset as alacritty does, and sends each title in order, since alacritty's own title events come after the tap's; one write and split writes post the same	2026-10-17 16:06:10 -04:00
ZacharyZhang-NY	52576bb8ee	T40: a press whose record went, a twin Linux stacked or one past the 64 kept, still shows its session; on Linux the body goes as markup's text; only a command C started posts its end, so a prompt bash 3.2 draws again posts nothing; the posting conditions are pure and tested; the decision says what macOS, Wayland and gpui's log do at a press and outside the bundle	2026-10-17 15:51:46 -04:00
ZacharyZhang-NY	efda50f142	T40: a session posts at each turn of its row to waiting, done or error, as the row's waiting mark is set, so one output posts as two would	2026-10-17 15:43:12 -04:00
ZacharyZhang-NY	3aaeb36bd1	T39: review round 2: a hidden session whose row turns to waiting marks it until shown, though it went on within the same output, so one output marks as two would; a change of herdr's records draws the view again, so an open hover follows them though the state holds	2026-10-17 15:41:39 -04:00
ZacharyZhang-NY	31de5e44ef	T40: sessions post system notifications through gpui's: each terminal posts its own in the order its output asks, when its row turns to waiting, done or error, when a command ends, and when a program asks through OSC 9, 777 or 99, only while its window is inactive or it is hidden, one every 3 seconds, titled with the session's name; a press goes to the window whose session posted it and shows it; OSC 99 hears untracked closes, alive lists what lives, in order before a later DA1, a close and w take one back where gpui can, a closed session's too, and the query says what holds where	2026-10-17 15:36:28 -04:00
ZacharyZhang-NY	53b8978dff	T39: review round 1: the view notes the state its row shows after each change it hears and tells the workspace of each, so a hidden session that waits, works and waits again in one output starts waiting again; hover reads the records and the last command's status, which the view now keeps, at each frame, stays under the pointer and scrolls within the window; the progress ring holds Ely's loader, since its track can match a selected row's ground	2026-10-17 15:12:29 -04:00
ZacharyZhang-NY	85bcbf466f	T39: a session's row shows its status, the most pressing first: waiting by its kind, an error, work with Ely's progress ring or a still loader, done, then a failed last command, then idle; hover lists each record's state, app, title and message, disarmed and cut; typing sees done and error, herdr's aside; a hidden session that starts waiting sets its title strong until shown	2026-10-17 14:43:56 -04:00
ZacharyZhang-NY	b504bfb9b8	T38: reviewed, ticked	2026-10-17 14:39:09 -04:00
ZacharyZhang-NY	7d56063df9	T38: review round 1: herdr's records take herdr: before the pane, a colon no OSC 7501 id holds, so a program's record named like theirs, herdr/build among them, stays through herdr's snapshots and its leaving	2026-10-17 14:24:42 -04:00
ZacharyZhang-NY	a175b18743	T38: inside herdr, which passes no OSC 7501 on, herdr's agents in a state it knows are the session's records, one a pane, their app, title and message from herdr's fields; a snapshot that lists other agents replaces them whole, and they go when herdr leaves the foreground	2026-10-17 11:51:56 -04:00
ZacharyZhang-NY	72acbe4bb0	T37: reviewed, ticked	2026-10-17 11:49:29 -04:00
ZacharyZhang-NY	628ed6b8c5	T37: review round 1: OSC 99 reads base64 group by group, so chunks split before encoding or after join, and the text, any UTF-8, is checked once whole; plain text alone refuses controls; a chunk that breaks a rule drops what its notification gathered; the tap keeps an OSC 99 body to its metadata and a full payload	2026-10-17 01:42:06 -04:00
ZacharyZhang-NY	ef33a0d34e	T37: programs ask for notifications through OSC 9, 777 and 99, read as Ghostty and kitty read them, OSC 99's chunks gathered by id with base64, options, close, query and alive, and the workspace hears each; OSC 9;4's progress stands as the root record until a program reports through OSC 7501, and again after a full reset	2026-10-17 01:23:09 -04:00
ZacharyZhang-NY	30e65e3435	T36: reviewed, ticked	2026-10-17 01:19:49 -04:00
ZacharyZhang-NY	4a269cb57a	T36: review round 1: every pair of a known key is checked before the last of each key wins, a clear's too, so a report that breaks a rule anywhere drops whole; a sequence counts its length from OSC through ST with the controls its body skips; the scanners pass DEL and high bytes by in an escape as vte does, so ESC DEL c resets the records too	2026-10-17 00:54:03 -04:00
ZacharyZhang-NY	42b272ece6	T36: programs report their status through OSC 7501, revision 0.4, checked as the spec checks them; each terminal keeps their records by id, replaced whole, cleared with those under them, 256 at most; a new prompt drops working, blocked and idle, a full reset all; the support query is answered with every state and kind, ahead of a DA1 sent after it; the pty's tests move to their own file	2026-10-17 00:46:07 -04:00
ZacharyZhang-NY	99db56bab1	T35: reviewed, ticked	2026-10-17 00:41:19 -04:00
ZacharyZhang-NY	11ec0effa0	T35: review round 3: a program's link target is kept whole, as the scanner keeps it, so a prompt inside a long link is numbered and gives the link back plain at its end, and the tap reads a 517th of its room to hold the growth; an ESC that ends an OSC the parser has goes on at once, so controls after it run where vte runs them	2026-10-17 00:26:34 -04:00
ZacharyZhang-NY	e2bb615a19	T35: review round 2: a program's link inside a prompt keeps its target and carries the prompt's number in its id, and a prompt that starts inside a link left open opens it again with the number, so a prompt all in links still starts on its first row; controls vte runs inside an escape go on before it, so only the ESC waits and a long run of them keeps the pty reading	2026-10-17 00:17:38 -04:00
ZacharyZhang-NY	310c219386	T35: review round 1: the tap holds an OSC that may be a mark until it ends and reads it as the session does, so 133;AB makes no prompt; each prompt's link carries the prompt's number, so a blank row within a prompt keeps one start; a program's own link keeps its cells and a prompt's link takes up again after it; controls vte runs in an escape go on before the link	2026-10-17 00:03:34 -04:00
ZacharyZhang-NY	15e83f5b7e	T35: Cmd-Up and Cmd-Down, Ctrl-Shift-Up and Ctrl-Shift-Down off macOS, jump to the prompt before or after the view's top, as Ghostty jumps: the tap rewrites each OSC 133 mark in place into an OSC 8 link of the terminal's own, which the parser keeps in the prompt's cells through scroll, reflow and history, and links pass over; the tap's OSC scanner ends a body as vte does	2026-10-16 23:46:58 -04:00
ZacharyZhang-NY	d3700f5278	T34: reviewed, ticked	2026-10-16 23:40:31 -04:00
ZacharyZhang-NY	2c25116703	T34: review round 3: PowerShell keeps a program's code unless the line surely ends in PowerShell, a cmdlet, a name not found or no command, so a call through a variable keeps it; bash without promptvars, zsh without prompt_percent and PowerShell without PSReadLine say at start what they cannot mark	2026-10-16 23:28:36 -04:00
ZacharyZhang-NY	d4bc6e69fa	T34: review round 2: bash 3.2 drops its DEBUG trap, which ran for key bindings, missed lone subshells and decided commands under extdebug, so it sends no C, and its D comes from the prompt by the command number; PowerShell takes a program's code only when the line's last pipeline ends in it, as its syntax tree and Get-Command name it; each check waits for the prompt before its line, so a repeated status is the line's own	2026-10-16 23:13:42 -04:00
ZacharyZhang-NY	0a22200725	T34: review round 1: PowerShell reads a native program's code only when one ran in the command, by the new object each run stores, and 1 for any other failure; bash 3.2 runs the user's own DEBUG trap after its own and keeps $_, and a line that is only a subshell sends its D from the prompt by the command number; zsh marks the first prompt under a theme that sets PS1 in its own hook, through sched; each shell run gets its own folder	2026-10-16 22:45:25 -04:00